...one of the most highly
regarded and expertly designed C++ library projects in the
world.
— Herb Sutter and Andrei
Alexandrescu, C++
Coding Standards
Normally when evaluating a function at say float
precision, maximal accuracy is assured by conducting the calculation at
double
precision internally,
and then rounding the result. There are two policies that control whether
internal promotion to a higher precision floating-point type takes place,
or not:
Policy |
Meaning |
---|---|
|
Indicates whether |
|
Indicates whether |
Suppose we want tgamma
to be evaluated without internal promotion to long
double
, then we could use:
#include <boost/math/special_functions/gamma.hpp> using namespace boost::math::policies; using boost::math::tgamma; // Define a new policy *not* internally promoting RealType to double: typedef policy< promote_double<false> > my_policy; // Call the function, applying the new policy: double t1 = tgamma(some_value, my_policy()); // Alternatively we could use helper function make_policy, // and concisely define everything at the call site: double t2 = tgamma(some_value, make_policy(promote_double<false>()));
Alternatively, suppose we want a distribution to perform calculations without
promoting float
to double
, then we could use:
#include <boost/math/distributions/normal.hpp> using boost::math::normal_distribution; using namespace boost::math::policies; // Define a policy: typedef policy< promote_float<false> > my_policy; // Define the new normal distribution using my_policy: typedef normal_distribution<float, my_policy> my_norm; // Get a quantile: float q = quantile(my_norm(), 0.05f);